Vacuum-assisted breast biopsies

Journal of Clinical Imaging 30 (2006) 99 – 107

Vacuum-assisted breast biopsies Experience at the Antoine Lacassagne Cancer Center (Nice, France) Claire ChapellierT, Catherine Balu-Maestro, Nicolas Amoretti, Chantal Chauvel, Isabelle Ben-Taarit, Isabelle Birtwisle-Peyrottes Centre Antoine Lacassagne, Unite´ de Radiologie Mammaire, 33 Avenue de Valombrose, 06189 Nice Cedex 2, France Received 20 June 2005; received in revised form 10 August 2005

Abstract This article presents the results of the first 318 aspiration-guided breast macrobiopsies performed using a dedicated table at the Antoine Lacassagne Cancer Center (Nice, France) between January 2001 and November 2002. A total of 288 procedures (90%) was performed for microcalcifications whereas that of 30 (10%) was performed for isolated opacities. There were 170 American College of Radiology (ACR) Category 4 (53.45%), 35 ACR 5 (11%), 111 ACR 3 (34.9%), and 2 ACR 2 lesions. There were 233 benign lesions that included 19 cases of atypical epithelial hyperplasia. Of the 85 malignant lesions, 33 had an invasive component. Seven of the atypical epithelial hyperplasia cases and all of the malignant lesions were managed surgically. Atypical epithelial hyperplasia was underestimated in 28.57% of the cases; ductal carcinomas in situ, in 21.15%. The positive predictive value of ACR 4 for the diagnosis of malignancy was 24.7% (42/128 cases), versus that of 12.6% for ACR 3 (14/97 cases). Among the 233 benign lesions, 128 were classified as ACR 4 and 6 as ACR 5 (all of these procedures obviated surgery). The advantages, drawbacks, and limitations of the technique are analyzed and indications are discussed, particularly for ACR 3 lesions. D 2006 Elsevier Inc. All rights reserved. Keywords: Breast biopsy; Microcalcifications; Large-core vacuum-assisted biopsy

1. Introduction The frequency and the severity of breast cancer have made this pathology a major public health problem. The development of screening programs, associated with progress in breast imaging techniques, has led to a tremendous increase in the number of subclinical radiological images. As the diagnosis of such images is often ambiguous, radiologists have had to become more and more efficient. Improvements in diagnostic accuracy are, to a great extent, due to the introduction of percutaneous sampling techniques. Fine-needle aspiration biopsy, extensively used in the 1990s, was progressively associated with, then replaced by, more aggressive percutaneous microbiopsy techniques that permit more accurate and complete histological study. T Corresponding author. Tel.: +33 4 02 03 12 65; fax: +33 4 92 03 10 09. E-mail address: [email protected] (C. Chapellier). 0899-7071/06/$ – see front matter D 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.clinimag.2005.08.002

Sample size and targeting accuracy rapidly conditioned the reliability of these techniques. Histopathologists were increasingly confronted with insufficient samples that complicated or even precluded diagnosis, in particular for borderline lesions such as atypical epithelial hyperplasia (AEH). These problems gradually led to the introduction of pecutaneous breast macrobiopsies thought to guarantee more reliable diagnosis owing to the size samples they provide.

2. Materials and methods 2.1. Materials The Fischer stereotactic imaging table and the vacuumassisted breast biopsy system used in this study were acquired by the Antoine Lacassagne Cancer Center (Nice, France) in January 2001.

100

C. Chapellier et al. / Journal of Clinical Imaging 30 (2006) 99 – 107

2.2. Methods Several steps were involved. At the initial appointment several days before the procedure, the radiologist and each patient discussed the radiology findings; the indication for the procedure and its feasibility were reviewed. If the procedure was feasible, it was explained to the patient, who was given an information sheet. This first consultation also permitted management of any problem related to allergies or blood coagulation disorders. The procedure was performed in the radiology department; each patient remained under supervision for an average of 30 min after the procedure had been completed.

In principle, the shortest possible approach was used, except for lesions in the upper internal quadrant, owing to the risk of scarring. The lateral arm approach was preferred for very small breasts and/or when the lesion lay close to the skin (b2 cm) or just behind the nipple. A 2- to 3-mm micro-incision was made following cleansing of the breast skin (Betadine preparation) and administration of local anesthesia with 1% lidocaine (usually 3–9 ml). The needle probe was inserted through the incision and positioned in the breast area containing the abnormality. The vacuum system was turned on (if the breast is not very thick, the probe can be prepared outside the breast and advanced manually) following stereotactic verification of the needle coordinates. Breast tissue was drawn into a sampling chamber where a rotating cutting device cut off tissue specimens. In general, two to three rotations of the probe’s thumbwheel sufficed; each rotation captured six to eight samples of tissue. At the end of the procedure, whenever a lesion had been completely removed, a small metal clip was placed in the biopsy site to facilitate future localization (Fig. 1). Manual pressure was applied to the biopsy site for 20 to 30 min by the radiologist or technician. Two postMammotome mammogram films of the breast were then taken to evaluate the modification of the lesion and clip positioning. In case of microcalcifications, the tissue samples were systematically radiographed. Two or three Steristrips were used to close the skin nick, covered by a pressure dressing. At the end of the procedure, patients were requested to fill out a questionnaire concerning tolerability. On average, the entire procedure, including manual compression, lasted 1 h. 2.3. Indications

Fig. 1. A, Mammogram: 5-mm cluster (ACR 4). B, Histology: no AEH. C, Post-Mammotome film: complete exeresis of the lesion (note the metal clip).

This prospective study covered 318 procedures performed from January 2001 to November 2002. The main indication for vacuum-assisted breast biopsies is small clusters of microcalcifications. Large clusters or small densities were also sampled but represented only a small portion of the study; most of such abnormalities were sampled at the start of the study, and analysis of such lesions was gradually abandoned. In all, 288 procedures (90%) were performed for microcalcifications and 30 (10%) for densities. Tissue samples were obtained for one or more lesions in a given patient during 1 or more procedures (17 patients underwent 2 procedures). Eligibility criteria were based on the BI-RADS categories established by the American College of Radiology (ACR). Eleven percent of the procedures (35 cases) concerned highly suspicious ACR 5 lesions. Most lesions were ACR 4 (n =170, 53.45%), but 111 (34.9%) were probably benign ACR 3 findings associated with risk factors (personal or family history of breast cancer, impossibility of surveillance, or before initiation of hormone replacement therapy).

C. Chapellier et al. / Journal of Clinical Imaging 30 (2006) 99 – 107

101

3. Results 3.1. Patients A total of 318 Mammotome procedures was performed in 301 patients, 192 of whom were postmenopausal (64%). The patients’ mean age was 56 years (range = 35 –78 years). Thirty-one procedures (10%) were performed using a lateral arm approach owing to the small size of the breast or the location of the lesion (10 lesions were located b2 cm from the skin and 5 were retroareolar). 3.2. Procedure tolerability Globally, tolerability was considered good by 269 patients (84.59%) and acceptable or poor by 49 (15.4%). In this category, 17 patients complained of intense cervical pain and 23 of pain that responded poorly to local anesthesia; 12 patients found the procedure stressful. 3.3. Technical difficulties Technical difficulties were encountered in 62 cases (19.4%) and were more or less all related to one another. In 22 instances (6.91%), the lesions were poorly visualized on the console. In 31 cases (10%), access to the lesions was difficult; in 24 (7.54%), the target was wrong. These technical difficulties were responsible for seven procedural failures (i.e., 7 cases in which the target was not reached). 3.4. Duration of the procedure

Fig. 2. A, Post-Mammotome study before repeat surgery (histology: lowgrade DCIS). B, Radiograph of the surgical specimen (note the clip, the harpoon, and several residual microcalcifications).

The entire duration of the procedure (including manual pressure application) was b1 h in 252 cases (79.24%). It lasted 90 min in 51 cases (16.03%) and 2 h in 15 (4.71%).

At the start of our experience, two ACR 2 abnormalities underwent biopsies: one patient with a fibroadenoma in the course of calcification who was scheduled for surgery anyway requested the procedure. The second ACR 2 abnormality corresponded to an intramammary node of recent appearance in a breast previously treated for cancer by subtotal mastectomy. The indications were also based on lesion size: 206 lesions were b1 cm (64.7%), 64 measured between 10 and 20 mm (20.1%), and 48 were N20 mm (15.9%).

3.5. Clip localization A small sterile clip was placed into the biopsy site at the end of the procedure in 211 cases (66.35%). The role of the clip is primordial. If repeat surgery is required, the clip facilitates localization of the biopsy site, which, in turn, influences the reliability of the surgical procedure [21] (Fig. 2). In 198 cases (93.8%), the position of the clip was considered satisfactory (i.e., it lay within 10 mm of the biopsy

Table 1 Target modification as a function of size Microcalcifications

b10 mm N10 mm Total

Densities

Disappearance

Partial excision

Total

Disappearance

Partial excision

Total

111 1 112

71 105 176

182 106 288

9 1 10

15 5 20

24 6 30

102

C. Chapellier et al. / Journal of Clinical Imaging 30 (2006) 99 – 107

Table 2 Target modification as a function of the number of thumbwheel rotations No. of rotations

Disappearance of microcalcifications

Disappearance of densities

V2 z3 Total

68 44 112

6 4 10

site). However, the distance was nearly always N7 mm (150 cases). This distance must obviously be recorded and taken into account in case of repeat surgery. In 13 cases (6.16%), the clip was located in an aberrant position (distance N 15 mm); this essentially concerned the D (Deep or Z) coordinate and was caused by the baccordeon phenomenon,Q as encountered during stereotactic localization. In 7 cases, the first clip remained lodged in the probe and a second one was required.

At the end of the procedure, two orthogonal analogical mammography films and the last digital film from the console (centered over the biopsy site) were used to evaluate lesion modification. Table 1 shows a list of modifications in the target as a function of size. The lesions that disappeared were b1 cm (111/112 clusters of microcalcifications and 9/10 densities). The table also shows that abnormalities N1 cm tended to persist. In contrast, as shown in Table 2, the number of tissue specimens obtained did not influence lesion persistence; they disappeared in the same manner after two or three thumbwheel rotations, probably because the microcalcifications were all concentrated in only a few samples.

Benign histology/ No. of cases

Malignant histology/ No. of casesa

Percentage of cancers

2 97 128 6 233

0 14 42 29 85

0% 12.6% 24.7% 82.8%

a Malignant lesions include, in a more or less associated manner, invasive ductal carcinomas, invasive lobular carcinomas, and carcinomas in situ.

There was no postprocedure abscess. Seven procedure failures occurred as the result of technical difficulties: –

– –

–

One axillary prolongation was inaccessible (ACR 5, microcalcifications); One breast was too small despite the lateral arm position (ACR 3, microcalcifications); Two clusters of microcalcifications had insufficient contrast, resulting in suboptimal targeting (ACR 4); Two clusters of microcalcifications had insufficient density (poor concentration), also resulting in suboptimal targeting (ACR 3); One low-contrast opacity measuring 10 mm (ACR 3) resulted in a targeting error. The ACR 4 and ACR 5 abnormalities underwent complementary surgical biopsies. Both cases proved to be malignant lesions [one high-grade ductal carcinoma in situ (DCIS) and one Grade 1 ductal carcinoma]; there were, thus, two true false-negative errors in our series (one ACR 5 and one ACR 4).

The four other procedure failures correspond to ACR 3 lesions that were all managed by close follow-up.

3.7. Complications and incidents A total of 123 subclinical hematomas was noted after the procedure (38.7%) and corresponded mammographically to round, well-delimited densities. In only 6 cases (1.8%) did secondary breast tension prompt ultrasound verification; the other hematomas all resorbed spontaneously.

ACR 3

ACR 4

ACR 5

Total

2

93

113

6

214

0 0 0 0

4 1 11 2

15 7 15 20

0 0 18 11

19 8 44 33

2

111

170

35

318

3.8. Histological findings Overall, 233 lesions were benign, including 19 cases of atypical epithelial hyperplasia (AEH) and 2 of atypical

Table 5 Histological correlations between vacuum-assisted breast biopsy and surgical excision

Table 3 Correlations of histological findings and ACR categories

Benign histology AEH DCIS + AEH Isolated DCIS Invasive carcinoma Total

ACR 2 ACR 3 ACR 4 ACR 5 Total (318)

–

3.6. Modification of the lesion

ACR 2

Table 4 Percentage of cancers as a function of ACR categories

233 benign lesions 85 malignant lesions

Surgery Biopsy: AEH Solitary in situ Invasive carcinoma in situ

No residual tumor

Benign

AEH

0

3

2

1

1

7

9

3

2

27

11

52

0 (complete excision)

1

0

11

21

33

In situ

Invasive carcinoma in situ

Total

C. Chapellier et al. / Journal of Clinical Imaging 30 (2006) 99 – 107

lobular hyperplasia (ALH). Of the 85 malignant lesions (26.7%), 33 had an invasive component. This distribution is conditioned by the aforementioned indications (Table 3). Table 4 shows the percentage of cancers as a function of ACR/BI-RADS categories. The ACR 5 category included 82 malignant lesions; the ACR 4, 24.7% malignancies; and the ACR 3, 12.6% malignancies. The one benign ACR 5 lesion corresponded to a fibroadenoma undergoing calcification. 3.9. Patient outcome Repeat surgery was performed for all of the carcinomas and for 7 of the 19 cases of atypical ductal hyperplasia (ADH). Radiological surveillance was offered for 12 of the 17 cases of AEH when the radiologically visible abnormality had been completely removed and/or the lesions were few in number and for all of the other benign lesions, including the 2 cases of ALH. Surveillance consisted in a follow-up mammogram 6–12 months after the Mammotome procedure. To date, no modification has been observed in any of the patients, but the follow-up period is still too short (just 4 months for the most recent patient). Annual follow-up of all of these patients must be continued to evaluate their outcome. Table 5 shows the histological correlations between vacuum-assisted breast biopsy and surgery for those patients who underwent a reoperation: –

– –

–

vacuum-assisted breast biopsy underestimated the lesions for 2 of the 7 patients with AEH who had a reoperation (28.5%); vacuum-assisted breast biopsy underestimated the nature of 11 of the 52 cases of solitary DCIS; among the 52 patients with DCIS, 12 had complete lesion excision by the Mammotome procedure; 38 had a residual tumor in the surgical specimen (73.07%); of the 33 invasive carcinomas (more or less associated with an in situ component), only 1 was completely removed by the Mammotome procedure; 32 patients had a residual tumor in the surgical specimen (97%);

103

4. Discussion 4.1. Indications The indications for vacuum-assisted breast biopsies must be discussed using a multidisciplinary approach (radiologist, surgeon, gynecologist, referring physician, histopathologist), taking medical context into account (family or personal history of breast cancer, degree of patient anxiety). The role of the radiologist is essential particularly for the evaluation of procedure feasibility and determination of a lesion’s ACR category. The main indications for percutaneous biopsies (and, thus, vacuum-assisted breast biopsies) are Category 4 abnormalities of the ACR BI-RADS classification. The positive predictive value (PPV) in such cases ranges from 5 to 70 for the diagnosis of malignancy, and percutaneous biopsy appears to be the most suitable for the situation [3,8,11]. Because nearly one of every two lesions in this category is benign, establishment of the histological diagnosis is essential to avoid excessive surveillance. However, this also means that one of every two lesions is malignant, and early diagnosis can help guarantee satisfactory management thereafter. In most series, ACR 3 abnormalities have a PPV V5 for malignancy [10,26,30]. This justifies mere surveillance, initially at close intervals then at longer intervals, because such lesions are probably benign [34]. When associated risk factors are present (personal or family history of breast cancer, impossible surveillance) or before the initiation of hormone replacement therapy, percutaneous biopsy is recommended [2]. For suspicious ACR lesions, vacuumassisted breast biopsy is a reliable alternative to open surgical biopsy and a valid substitute for extemporaneous examination that can be difficult in case of microcalcifications and may give a false-negative error. Prior diagnosis allows planning of the surgical procedure and, if necessary, use of the sentinel lymph node technique. 4.2. Advantages The advantages of vacuum-assisted breast biopsy include its cost (less expensive than open surgical biopsy),

Table 6 Vacuum-assisted breast biopsies (8 or 11 gauge): underestimation of AEH Authors

Needle (gauge)

Type of lesion

No. of AEH/No. of biopsies

Underestimation (%)

Heywang-Kobrunner et al. [15] Liberman et al. [22] Hagay et al. [14] Philpotts et al. [31] Darling et al. [10] Brem et al. [7] Mayras and Doutriaux-Dumoulin [27]

11 11 11 11 11 11 11 8 11

Microcalcification Microcalcification Microcalcification Microcalcification Microcalcification Microcalcification Microcalcification Microcalcification Microcalcification

4/134 10/112 27/300

0 10 4.7 26.7 20 25 0/4 2/4 2/7 (28.5)

Chapellier et al. (present study)

16/82 18 4/140 8/66 7/158

104

C. Chapellier et al. / Journal of Clinical Imaging 30 (2006) 99 – 107

tolerability, diagnostic reliability, and absence of scarring effects [12,24]. One of the major advantages of percutaneous biopsies and in particular of vacuum-assisted breast biopsies is improvement of the PPV of open surgical biopsies that can result in fewer surgical operations for benign lesions. Like other authors [4,9], we were able to avoid reoperation for numerous benign ACR 4 and ACR 5 abnormalities (128/170 ACR 4 lesions and 6/35 ACR 5 lesions). The reliability of vacuum-assisted breast biopsies is identical to that of open surgical biopsies, and their false-negative rates are similar [17,24]. There were four false-negative errors in our series: two cases of AEH (one DCIS and one invasive ductal carcinoma after surgery) and two procedure failures (one each in ACR 4 and ACR 5). Procedure tolerability was excellent; complaints were minor and related essentially to cervical neck pain owing to the prone position. The psychological status of patients during the postprocedure period appears to be linked to the benign or malignant nature of the lesions (more traumatic reaction in case of a malignant lesion), the information given, and the attitude of the medical staff [3]. Complications are rare and correspond essentially to bruising that usually resolves spontaneously and does not require any treatment [24].

probe at vacuum-assisted breast biopsy). Nevertheless, the risk persists. Underestimation of AEH and DCIS typically occurs in 18 – 88% (average= 50%) of cases in microbiopsy series performed with 14-gauge needles (17%); it continues to be a problem with vacuum-assisted breast biopsies, occurring in an estimated 0 –26% of cases (Table 6). The record level of 28.6% noted in our series was probably related to the small number of cases [10,14,15,25,27,31]. Underestimation is directly related to the extent of lesion excision [1,24,31]. Excision by vacuum-assisted breast biopsy is all the more complete when lesions are small. This was confirmed in our study, where lesions b1 cm were usually completely removed. In contrast to lesion size, the number of thumbwheel rotations had no influence on results (Tables 3 and 4). For many authors, such persistent underestimation mandates systematic repeat surgery whenever AEH is discovered by vacuum-assisted breast biopsy [18,24]. However, this systematic approach remains controversial. Several authors have attempted to identify cases of AEH requiring repeat surgery and those that can merely be kept under surveillance. Adrales et al. [1] reoperated on 62 patients found to have ADH at Mammotome biopsy; ADH was underestimated in 15 of the cases (n = 9 patients) as malignancy was found at subsequent excision. In that study, several variables predictive of an association of AEH and cancer had a combined sensitivity of 100 and a specificity of 80: markedly atypical hyperplasia, incomplete removal of microcalcifications, personal history of contralateral breast cancer, and family history of breast cancer. Reoperation is advisable for these, more or less, associated situations. In other cases, after complete lesion excision, surveillance can be proposed every 6 months, then annually. This is the approach we adopted for 12 of the 19 cases of AEH in our study that corresponded to all of these criteria. To date, no modification has been noted on the post-Mammotome procedure mammograms obtained for these patients. In terms of underestimation of DCIS and invasive carcinoma, for AEH, the introduction of vacuum-assisted breast biopsies has reduced the rate of underestimation of DCIS and invasive carcinoma [7,17] that classically varies from 15 to 67 in microbiopsy series performed with 14gauge needles [10,21,27,31]. With vacuum-assisted breast biopsies, underestimation continues to be a problem in 6–20% of cases (Table 7); it

4.3. Drawbacks The drawbacks of vacuum-assisted breast biopsy are mainly related to lesion management. Certain situations such as histological underestimation of disease are already familiar as they are encountered with microbiopsy techniques. Other situations are new and related to the technique itself: this is the case for management of ACR 3 lesions and lobular neoplasia. Histological underestimation occurs in situations where percutaneous biopsy detects AEH that subsequent open surgical biopsy proves to be associated with DCIS or, less often, an invasive carcinoma. In other cases, a DCIS diagnosed at percutaneous biopsy is found to be associated with an invasive component after open surgical biopsy. The frequency of such underestimation has unquestionably decreased since large-dimension percutaneous specimens have become available [6,10,16] (from a mean of 17 mg for a specimen obtained with a 14-gauge automatic instrument to that of 94 mg for a specimen obtained with an 11-gauge Table 7 Vacuum-assisted breast biopsies (8 or 11 gauge): underestimation of DCIS Authors

Needle (gauge)

Type of lesion

No. of procedures

Underestimation [n (%)]

Won et al. [35] Liberman et al. [22] Darling et al. [10] Philpotts et al. [31] Mayras and Doutriaux-Dumoulin [27]

11 11 11 11 11 8 11

Microcalcification Microcalcification Microcalcification Microcalcification Microcalcification Microcalcification Microcalcification

20/236 28 18/73 48/86 36/173 16/66 14/158

3/20 4/28 10 9/48 7/36 1/16 5/14

Chapellier et al. (present study)

(15) (14) (18) (20) (6) (35.7)

C. Chapellier et al. / Journal of Clinical Imaging 30 (2006) 99 – 107

105

reached a record level of 35.7% in our small study population. In practice, underestimation of the invasive nature of a lesion is less important in terms of diagnosis because discovery of a DCIS at vacuum-assisted breast biopsy is always an indication for surgical excision. 4.4. New situations 4.4.1. Management of lobular neoplasia Lobular neoplasia includes lobular carcinomas in situ (LCIS) and ALH. These diffuse, bilateral lesions are usually too small to be seen radiologically and are often associated with other histological lesions. Owing to our lack of experience with the technique and the paucity of information in the literature, lobular neoplasia involves a problem for management after discovery (often fortuitous) in the surgical specimen or vacuum-assisted breast biopsy specimen. The question is whether reoperation is required for complete lesion removal or whether radiological surveillance will suffice. Lobular carcinomas in situ have long been considered a benign lesion [23], yet 20% of all cases become invasive and both the homolateral and contralateral breasts are at risk. Moreover, LCIS can also undergo transformation into an invasive lobular carcinoma or an invasive ductal carcinoma, sometimes after a very long period of latency (up to 20 years). Increasingly, LCIS are considered precancerous lesions that must be treated. Certain authors [32 –34] recommend reoperation after any diagnosis of LCIS. Liberman et al. [21] and Jacobs et al. [18], however, Fig. 4. A, Five-millimeter round cluster of punctate microcalcifications (ACR 3). B, Radiograph of the tissue specimens (histology: lowgrade DCIS).

felt that close surveillance suffices, except in three situations where surgical excision is preferable: – – –

Fig. 3. A, Mammogram: 5-mm cluster of a punctate and (dust-like) microcalcifications (ACR 3). B, Radiograph of the tissue specimens (histology: low-grade ductal carcinoma in situ).

association with a high-risk lesion such as AEH; association with radiating scar lesions; and discrepancy between radiological and histological findings (histological lesions of LCIS for densities or suspicious effects of convergence classified as ACR 4 or ACR 5).

No case of solitary LCIS or AEH was encountered in our series. In case of association with an invasive carcinoma or DCIS, reoperation was always mandated by the latter lesions anyway. The histological limit between ALH and LCIS remains unclear, and we do not know whether they are a continuation of a same pathology or two distinct lesions with different courses. Owing to our lack of experience and the limited information available on the subject, the attitude to be adopted when confronted with such lesions remains to

106

C. Chapellier et al. / Journal of Clinical Imaging 30 (2006) 99 – 107

be defined. The two cases of ALH in our series (both benign lesions) were managed solely by surveillance, and follow-up to date has been reassuring. 4.4.2. Management of ACR 3 lesions American College of Radiology Class 3 lesions have a PPV for malignancy of b5% [10,23,31]. The Agence Nationale d’Accre´ditation et d’Evaluation en Sante´ [2] thus recommends mere surveillance, initially at close intervals then at longer intervals; percutaneous biopsy is recommended for patients with associated risk factors. We were surprised to find a high rate of cancer in our ACR 3 group (12.6; four cases of DCIS, including one with an invasive component at reoperation). Aware of the difficulty in classifying microcalcifications in category ACR 3 and the potential variability in this evaluation [5,19,20,28,30], we had these files reinterpreted by a second and sometimes even a third radiologist. Reanalysis confirmed the 12.6% rate of cancers in our ACR 3 group (Figs. 3 and 4). This high level has also been encountered by other authors such as Se´ror et al. [33] (in 87 patients) and Mayras and Doutriaux-Dumoulin [27] (in 17 patients), who all found a PPV close to 20. If this high PPV is confirmed in other series, systematic biopsy might be advisable for all ACR 3 lesions. Although this strategy remains to be validated, numerous authors have already emphasized the value of percutaneous biopsy for such lesions, rather than surveillance, which can involve considerable constraints [13,27,29,33].

5. Conclusion Our initial experience at the Antoine Lacassagne Cancer Center with 318 vacuum-assisted breast biopsy procedures confirms the findings of other teams. Owing to its high reliability, this well-proven technique allows complete removal or tissue sampling of subclinical mammary lesions under excellent conditions of patient comfort and tolerability. The indications for the procedure must be discussed in an interdisciplinary manner, with selection being based on the ACR classification. The main indications are ACR 4 and at-risk ACR 3 abnormalities, but, increasingly, the technique is being extended to no-risk ACR 3 findings and ACR 5 lesions. The risk of underestimation is reduced for small (V1 cm) lesions. For large lesions not amenable to complete percutaneous excision, we recommend mere surveillance for benign abnormalities and reoperation for AEH. The optimal approach for LCIS and ALH remains to be determined.

References [1] Adrales G, Turk P, Wallace T, Bird R, Norton HJ, Greene F. Is surgical excision necessary for atypical ductal hyperplasia of the breast diagnosed by Mammotome? Am Surg 2000;180:313 – 5.

[2] Agence Nationale d’Accre´ditation et d’Evaluation en Sante´. Conduite a` tenir diagnostic devant une image mammographique infra-clinique anormale. ANAES. Recommandations pour la pratique clinique. Paris 1998. [3] Barreau B, Dilhuydy MH, Henrique`s C, Gilles R, Valentin F. Imagerie interventionnelle: indications, re´sultats et e´valuation de la qualite´. XXIIIemes journe´es Nationales de la Socie´te´ Franc¸aise de Se´nologie et de Pathologie Mammaire. Strasbourg. 10 –12 Octobre 2001. [4] Beck RM, Gotz L, Heywang-Kobrunner SH. Stereotaxic vacuum core breast biopsy—experience of 560 patients. Swiss Surg 2000;6: 108 – 10. [5] Berg WA, Campassi C, Langenberg P, Sexton MJ. Breast imaging reporting and data system: inter- and intraobserver variability in feature analysis and final assessment. AJR Am J Roentgenol 2000;174:1769 – 77. [6] Berg WA, Krebs TL, Campassi C, Magder LS, Sun CC. Evaluation of 14 and 11 gauge directional, vacuum-assisted biopsy probes and 14 gauge biopsy guns in a breast parenchymal model. Radiology 1997;205:203 – 8. [7] Brem RF, Behrndt VS, Sanow L, Gatewood OM. Atypical ductal hyperplasia: histologic underestimation of carcinoma in tissue harvested from impalpable breast lesions using II gauge stereotactically guided directional vacuum-assisted biopsy. AJR Am J Roentgenol 1999;172:1405 – 7. [8] Burbank F. Stereotactic breast biopsy: comparison of 14- and 11-gauge Mammotome probe performance and complication rates. Am Surg 1997;63:988 – 95. [9] Cangiarella J, Waisman J, Symmans WF, Gross J, Cohen JM, et al. Mammotome core biopsy for mammary microcalcifications. Cancer 2001;91:173 – 7. [10] Darling ML, Smith DN, Lester SC, et al. Malignant mammographic lesions at stereotactic vacuum-assisted biopsy. Radiology 1998;206: 711 – 5. [11] Dilhuydy MH, Henrique`s C, Barreau B, Palussie`re J, Joyeux P, Bussie`res E. Biopsies chirurgicales percutane´es ABBI et microbiopsies 8 gauge assiste´es par le vide MIBB. L’expe´rience multidisciplinaire du Centre de Lutte Contre le Cancer de Bordeaux. J Le sein 1999;9:118 – 25. [12] Feillel V, Lafaye C, Le BouJdec G, et al. System ABBI: expe´rience franc¸aise. 22e`mes journe´es de la socie´te´ franc¸aise de se´nologie et de pathologie mammaire. Toulouse, 12–14 Octobre 2000. [13] Geller BM, Barlow WE, Ballard-Barbash R, et al. Use of the American College of Radiology BI-RADS to repeat the mammographic evaluation of women with signs and symptoms of breast disease. Radiology 2000;222:536 – 42. [14] Hagay C, Che´rel P, Becette V, Garbay JR. Evolution des biopsies ste´re´otaxiques des le´sions mammaires non palpables. Premiers re´sultats sur table de´die´e nume´rise´e avec aiguilles de 14 G. Gynecol Obstet 1999;6:1 – 6. [15] Heywang-Kobrunner SH, Schaumioffel U, Viehweg P, Hofer H, Buchmann J, Lampe D. Minimally invasive stereotaxic vacuum core breast biopsy. Eur Radiol 1998;8:377 – 85. [16] Jackman RJ, Burbank F, Parker S, et al. Atypical ductal hyperplasia diagnosed at stereotactic breast biopsy: improved reliability with 14-gauge, directional, vacuum-assisted biopsy. Radiology 1997;204: 485 – 8. [17] Jackman RJ, Nowels KW, Rodriguez J, et al. Stereotactic, automated, large-core needle biopsy of non palpable breast lesions: false-negative and histologic underestimation rates after long-term follow-up. Radiology 1999;210:779 – 805. [18] Jacobs TW, Connolly JL, Schnitt SJ. Nonmalignant lesions in breast core needle biopsies. Am J Surg Pathol 2002;26:1095 – 110. [19] Kerlikowske K. Timeliness of follow-up after abnormal screening mammography. Breast Res Treat 1996;40:53 – 64. [20] Kim SJ, Morris EA, Liberman L, et al. Observer variability and applicability of BI-RADS terminobreast MR imaging: invasive carcinomas as focal masses. AJR Am J Roentgenol 2001;177(3):551 – 7.

C. Chapellier et al. / Journal of Clinical Imaging 30 (2006) 99 – 107 [21] Liberman L, Dershaw DD, Morris EA, Abramson AF, Thornton CM, Rosen PP. Clip placement after stereotactic vacuum-assisted breast biopsy. Radiology 1997;205:417 – 22. [22] Liberman L, Dershaw DD, Rosen PP, Morris EA, Abramson AF, Borgen PI. Percutaneous removal of malignant mammographic lesions at stereotactic vacuum-assisted biopsy. Radiology 1998;206: 711 – 5. [23] Liberman L, Sama M, Susnik B, et al. Lobular carcinoma in situ at percutaneous breast biopsy: surgical biopsy finding. AJR Am J Roentgenol 1999;173:291 – 9. [24] Liberman L, Sama Michelle P. Cost–effectiveness of stereotactic 11-gauge directional vacuum-assisted breast biopsy. AJR Am J Roentgenol 2000;175:53 – 8. [25] Liberman L. Percutaneous imaging-guided core breast biopsy: state of the art at the millenium. AJR Am J Roentgenol 2000;174: 1191 – 9. [26] Lo JY, Markey MK, Baker JA, Floyd C. Cross-institutional evaluation of BI-RADS predictive mode mammographic diagnosis of breast cancer. AJR Am J Roentgenol 2002;178:457 – 63. [27] Mayras C, Doutriaux-Dumoulin I. Biopsie mammaire assiste´e par aspiration L’expe´rience des CRLCC Paul Papin (Angers) et Rene´ Gauducheau (Nantes). Journal Le Sein 2001;12:133 – 43. [28] McKay C, Hart CL, Erbacher G. Objectivity and accuracy of mammogram interpretation using the BI-RADS final assessment

[29]

[30] [31]

[32]

[33]

[34] [35]

107

categories in 40- to 49-year-old women. J Am Osteopath Assoc 2000;100:615 – 20. Obenauer S, Fisher U, Baum F, Dammert S, Fuzesi L, Grabbe E. Stereotactic vacuum core biopsy of clustered microcalcific classified as BI-RADS 3. Rofo 2001;173(8):696 – 701. Orel SG, Kay N, Reynoids C, Sullivan DC. BI-RADS categorization as a predictor of malignancy. Radiology 1999;211:845 – 50. Philpotts L, Shaheen N, Carter D, Lange R, Lee C. Comparison of rebiopsy rates after stereotactic core needle biopsy of the breast with 11-gauge vacuum suction probe versus 14-gauge needle and automatic gun. AJR Am J Roentgenol 1999;172:683 – 7. Reynoids H. Core needle biopsy of challenging benign breast conditions: a comprehensive literature review. AJR Am J Roentgenol 2000;174:1245 – 50. Se´ror JY, Antoine M, Scetbon F, et al. Apport des macrobiopsies ste´re´otaxiques par aspiration dans la prise en charge des microcalcifications mammaires: premie`re se´rie prospective de 115 cas. Gynecol Obstet Fertil 2000;28:804 – 19. Sickles EA. Management of probably benign breast lesions. Radiol Clin North Am 1995;33:1123 – 330. Won B, Reynoids H, Lazaridis C, et al. Stereotactic biopsy of ductal carcinoma in situ of the breast using an 11-gauge vacuum-assisted device: persistent underestimation of disease. AJR Am J Roentgenol 1999;173:227 – 9.